Blog: What You See as Wage Growth May Be Noise

So much depends on wage growth these days.  The over 10bps backup in rates has largely been attributed to better economic conditions and therefore anticipation of further rate hikes. Rightly or wrongly, inflation pressures are on investor’s minds.

One of the claims last week by the popular press is the “Wage Growth Fastest Since 2009 (CNN)” claim that average hourly earnings has gone up a lot.  If you look at the chart below it IS REALLY close to September 2017 level...  There are lot of issues with looking a economic data on a year-over-year basis (but that’s a topic to be discussed elsewhere). 

Ignoring that, lets look under the covers of the report to see if we think this move is noise or there’s an emerging trend.  (HINT: unless you work in information sector, it's mostly noise).  

Expect some of this strength to recede on March 9th when the next hourly earnings report is released.


I’m not the first person to discuss this.  The WSJ has a great chartpack poking holes at the strength of the report (WSJ). Even the Fed’s Kashkari says he doesn’t see strength yet (WSJ).

I’m going to do a little better than opinion here. We can take the top 8 sectors in the labor market, create forecasts for each of them and aggregate them up to create a forecast for Total Private Hourly Earnings.  

The take-away is to expect ~0.22% m/m increase in next month’s average hourly earnings growth. This is below last month’s 0.3% rise and the y/y chart will DECLINE to 2.78% from 2.89%.


Over the next three month’s there’s a lot of uncertainty, but the bias in the next release is for slower growth


Construction has been a strong positive contributor lately, however it is highly volatile.  We expect information jobs to continue contributing the largest wage growth of any sector.   Mining and Logging we expect to be the smallest contributor.

Here’s a table of the contributions by sector along with sector weights.  All changes are in monthly log changes.

Component Forecasts returns vol wt return contrib. vol


Duration of Trend (Months)
mining and logging 0.14% 0.36% 1.0% 0.001% 0.04% 22.8
manufacturing 0.17% 0.21% 15.9% 0.028% 0.08% 25.0
trade 0.20% 0.25% 29.4% 0.058% 0.13% 36.1
professional and business services 0.23% 0.25% 23.2% 0.054% 0.12% 23.9
other services 0.25% 0.19% 5.7% 0.014% 0.04% 26.5
leisure and hospitality 0.25% 0.27% 13.1% 0.032% 0.10% 10.6
construction 0.25% 0.54% 8.6% 0.022% 0.16% 9.4
information 0.33% 0.35% 3.1% 0.010% 0.06% 24.3

As we can see information has the greatest contribution.  However professional and business services has the largest weight


The thing to really note here is the amount of month-month volatility in these sub-components.  Professional and Business Services was strong last month (see below) and our Breakpoint© models currently caution against extrapolating that move.


The green lines above signal our current 1 standard deviation estimates, and the black line is the mean forecast.  The red blob is the Breakpoint probability.  This is the estimation of the location of the latest change in trend or volatility.  

The key insight here is that the component with one of the largest weights had a good month last month, but NOT ENOUGH to accelerate the trend.

  • Do you think a weaker wage report on March 9th, 2019 would be enough to stop the sell-off in rates?
  • Do you think I’m missing anything in this analysis?

Let me know in the comments below or email me at

Full forecasts for Non-Farm Payrolls, Average Hourly Earnings, and a suite of other key macro indicators are available to clients of Astrocyte only.  Please let me know if you have interest.

If you liked this analysis, we also did a deeper dive on Business Cycles a little while back: Using Machine Learning to Redefine Business Cycles.

Thanks for geeking out with me,

Sean Kruzel

econSean Kruzelwages, econ, us, blsComment